Extract Audio Online — Free, Fast, Private

How Extract Audio Works

Every video file is a container — MP4, WebM, MOV — holding at least two separate streams multiplexed together: a video stream and an audio stream, each encoded with its own codec. Technically, pulling the audio out is a demuxing operation: separating the streams without touching the encoded data inside either one. Desktop tools that use FFmpeg do exactly that — they copy the audio bytes straight out of the container, bit for bit, which is fast and lossless. Browsers don't expose that kind of low-level container access to a web page, so this tool takes a different route: it plays the video silently in the background, routes the decoded audio through the browser's recording pipeline, and captures the result as a new file. The output is a faithful copy of what you'd hear, even if it isn't a byte-identical extraction of the original stream.

That approach has a practical implication worth knowing: because the tool has to actually play the video to capture its sound, extraction takes roughly as long as the video's runtime, not an instant flash like some server-side tools promise. A three-minute clip takes about three minutes to process. In exchange, nothing gets uploaded anywhere — the video is decoded, played, and recorded entirely inside your browser tab using the Canvas and MediaRecorder APIs, the same building blocks browsers use for screen recording and video calls.

People reach for this tool in a handful of recurring situations: turning a recorded Zoom call or webinar into a podcast-style audio file, ripping the narration off a screen-capture tutorial, pulling a voiceover or sound effect out of a longer edit to reuse elsewhere, or simply converting a video someone sent you into something you can listen to without staring at a screen. It sits alongside the rest of the /video-tools/ toolkit as the inverse of Mute Video, which strips audio out and keeps the picture — Extract Audio keeps the sound and discards the picture. If your video's file size is the real problem rather than needing the audio separately, Compress Video is the better starting point.

One nuance worth flagging: the extracted WAV is a re-recording, not a direct stream copy, so extremely precise audio-restoration work (forensic analysis, exact loudness matching against the source) may show tiny artifacts from the re-encode pass. For everyday use — podcasts, voice notes, sampling, archiving — the difference is not audible. If you need the file in a smaller, more shareable format afterward, run the WAV through an MP3 or other audio converter as a second step.